# Mathieu Haye **Data, AI & Finance.** Builder. Markets-obsessed. Data & AI-driven. Paris-based. AFD-bound. 22 years old. Paris region, France. ALM apprentice at AFD (Agence Française de Développement) from September 2026. Incoming student at the MSc Data & AI for Finance, Albert School x Mines Paris-PSL (2026-2028). Freelance consultant in CRM, data engineering and applied AI since October 2025. - **Email**: contact@mathieuhaye.fr - **Phone**: +33 6 61 51 32 89 - **LinkedIn**: https://www.linkedin.com/in/mathieu-haye/ - **Languages**: French (native), English (B2/C1) --- ## Profile I build trading bots, data pipelines and AI workflows that turn market data into decisions. Freelance consultant since 2025, active investor across ETFs, crypto and real estate. Dual background in **Economics & Management** (Université Gustave Eiffel) and **Digital Innovation** (L3 MITIC), now heading into the **MSc Data & AI for Finance** at Albert School x Mines Paris-PSL. I build end-to-end data products: trading bots, scoring pipelines, decision dashboards, AI workflows. I invest actively in ETFs, crypto and real estate, and I model what I buy. The next step is to add the quantitative foundation that makes all of this rigorous. ### Skills - **Finance & Quant**: Bloomberg Terminal (BMC), ALM basics, fixed income, portfolio management, ETF / crypto, real estate modeling. - **Python & Data**: pandas, numpy, scipy, scikit-learn, SQL, ETL, APIs, scraping. Power BI, Looker Studio. - **AI & Automation**: Claude API, prompt engineering, AI scoring, n8n workflows (93+ nodes), Make, Apps Script. - **CRM & Platforms**: Salesforce (Admin, Apex, SOQL, LWC), Pipedrive, HubSpot, Brevo, Odoo. --- ## Projects ### 1. Crypto Trading Bot (MEXC) Mean reversion on BTC and ETH with RSI + Bollinger Bands entries. Paper ran +3.2% over 90 days. Live pilot ended at -2.1%: taker fees plus slippage ate the edge. Next iteration on maker-rebate orders. - **Strategy**: Mean reversion, BTC + ETH - **Paper P&L**: +3.2% over 90 days - **Live P&L**: -2.1% over 30 days - **Win rate**: 46% (214 trades) - **Stack**: Python, MEXC API, pandas, numpy, Streamlit, TA-Lib **Key lesson**: never trust a backtest that does not model friction costs tick by tick. Fees alone explained ~4 percentage points of the paper-live gap. ### 2. Bloomberg-Style Dashboard Personal multi-asset tracker (crypto, stocks, ETF) with technical indicators, AI-generated commentary on market open/close, DCA tracker and Telegram alerts. - **Scope**: 6 assets (BTC, ETH, SOL, NVDA, TTE, CW8) - **AI**: Claude Haiku 4.5 - **Alerts**: Telegram, 3 tiers (urgent, daily digest, weekly review) - **Schedule**: 08:50 / 13:00 / 17:30 Paris - **Stack**: Next.js 14, TypeScript, Claude API, Recharts, CoinGecko, Yahoo Finance Features: scheduled AI analysis, technicals (RSI, MACD, Bollinger, ATR on 1D/1W/1M/3M), What-If Monte Carlo simulator, Pearson correlation heatmap (6x6), DCA tracker with cost basis, events calendar (earnings, dividends, ECB, Fed), trade journal. ### 3. IchimokuSignal Pro (Backtested) TradingView indicator for long-term stock picking: bounce detection, trailing 15% stop loss, higher timeframe candles. Python backtest reproduces the Pine logic with commissions and slippage. - **Version**: v3.3 (Bounce + 15% SL) - **Net return (AAOI, 5y)**: +187% vs +1088% Buy & Hold - **Win rate / PF / MaxDD**: 54% / 1.62 / 24% - **Stack**: Pine Script v6, TradingView, Python, yfinance, pandas, numpy Honest take: the strategy loses to Buy & Hold in absolute terms but cuts drawdown by more than half. Single-stock backtest with overfit risk, forward-testing on a 15-ticker basket is next. ### 4. Real Estate Investment Model Python tooling to compare SCI / IS vs LMNP regimes, model mortgage scenarios, and rank French cities on a rental yield index. Used to size my own portfolio. - **Scope**: SCI/IS vs LMNP, mortgage simulation - **Data**: INSEE, notary stats, rental yields - **Cities ranked**: 34 French metros - **Stack**: Python, pandas, Notion, INSEE open data, Google Sheets Components: regime comparator (IRR 10y), mortgage simulator, city index (yield / vacancy / price momentum / notarial tax), macro overlay. ### 5. IA Brew, AI Newsletter Fully automated weekly newsletter: ingest from 20+ sources, AI scoring and summarisation, HTML rendering, delivery via Brevo. Runs on autopilot, no human in the loop. - **Type**: Automated newsletter - **Workflow**: 93+ nodes n8n - **Frequency**: Weekly, unattended - **Stack**: n8n, Claude API, Apify, Brevo, HTML templating Pipeline stages: ingest -> dedupe (content fingerprinting) -> score & summarise (Claude) -> render -> send -> observability (Google Sheets logs). ### 6. Multi-source Job Scorer + ATS PDF Generator Aggregates WTTJ, JobTeaser and LinkedIn via APIs, applies a weighted scoring algorithm against my profile. For each high-scoring offer, auto-generates an ATS-optimised CV and cover letter in PDF, tailored to the job description. - **Sources**: WTTJ, JobTeaser, LinkedIn - **Jobs scored**: 240+ - **Auto-generated**: ATS CV + cover letter in PDF - **Stack**: Python, Claude API, ReportLab, HTML/JS, SQLite ATS rules respected by construction: single-column layout, no tables, no images behind text, real font files, standard section names, machine-parseable headings. Parses correctly on Workday, Greenhouse, Lever. ### 7. e-Enfance / 3018 Platform Salesforce build-out for the 3018 child protection hotline: Apex REST APIs, Lightning Web Components, 3CX telephony integration, Einstein Bot triage and CSP rules. - **Client**: e-Enfance / 3018 - **Scope**: Full platform customisation - **Integrations**: 3CX, Einstein Bot, Web-to-Case - **Role**: Freelance consultant --- ## Background ### Sep 2026 → 2028 (incoming): ALM Pilot Apprentice **AFD - Agence Française de Développement, Paris** Two-year apprenticeship inside the ALM team of a public development bank, supervised by a senior ALM financial pilot. Paired with the MSc Data & AI for Finance on a 4 / 1 days rhythm. - **ALM strategies**: contribute to designing and deploying the strategies that cap the Group's interest rate and liquidity risks. - **Disbursement models**: update the ALM models that forecast loan disbursement profiles. - **Metric pipeline**: improve the production and reliability of the core ALM metrics. - **ALCO reporting**: help draft sections of the Asset-Liability Committee reports. - **Stack**: VBA, Python, ALM financial IS. ### 2026 → 2028 (expected): MSc Data & AI for Finance **Albert School x Mines Paris-PSL** Two-year MSc built on three pillars: - **Quantitative modeling**: advanced statistics, machine learning, risk analysis, interest rate and liquidity models. - **Applied data science**: advanced SQL, Python (pandas, numpy, scipy, scikit-learn), Power BI, decision dashboards. - **Financial markets**: ALM, fixed income products, banking regulation (Basel III / IV), market microstructure. Goal: turn a self-taught builder into a rigorous quant, ready to ship on a real ALM desk from day one, the apprenticeship at AFD. ### 2026: Bloomberg Market Concepts (BMC) Certification covering economic indicators, currencies, fixed income, interest rate risk and equities. ### Oct 2025 → present: Freelance Consultant **Clients: e-Enfance / 3018, Fromagerie Ermitage, Horus Condition Report, Profile Club.** - **Salesforce build, e-Enfance / 3018**: full platform customisation, Apex REST, Lightning Web Components, 3CX telephony, Einstein Bot triage, Web-to-Case, CSP rules. - **Data monitoring, Fromagerie Ermitage**: 93-node n8n workflow for automated press and social media monitoring. 19-indicator keyword scoring, temporal filtering, weekly reports auto-generated. - **CRM restructuring, Horus Condition Report**: Pipedrive migration and bilingual (FR / EN) sales automations. - **Data & analytics, Profile Club**: 146-record member database, cohort analysis, campaign segmentation, KPI dashboards on Google Apps Script. ### Sep 2024 → Sep 2025: Digital Project Coordinator **Concilium, Paris.** Digital project management agency, 150+ projects / year. - Project coordination: backlog management, steering committee prep, deliverable tracking, client reporting across 150+ projects. - CRM admin: OHME and Pipedrive deployment, contact data structuring, segmentation, campaign exports. - Automated reporting: sector monitoring newsletter and internal AI newsletter built on n8n + Brevo. ### Education (continued) - **2024 → 2025**: BSc Digital Innovation & IT (L3 MITIC). Université Gustave Eiffel, Serris. - **2022 → 2024**: BSc Economics & Management (L1-L2). Université Gustave Eiffel, Serris. - **2021 → 2022**: BSc Computer Science (L1). ESGI, Paris. --- ## MSc application ### What I bring, what I need For two years I have been building what I could on my own: trading bots, backtest frameworks, automated research pipelines, a real estate investment model for my own portfolio. The outcome is a working stack. The limit is obvious too: I can ship, but I lack the quantitative depth to trust my own numbers. The **MSc Data & AI for Finance** is the exact intersection I need. **Mines Paris-PSL** for the quantitative rigor (ML, risk modeling, ALM, Basel III/IV). **Albert School** for the business reflex and the access to the finance industry. Together they turn a self-taught operator into a serious quant profile. Goal after the MSc: continue to build the decision tools I already prototype today, but with the science behind them. The ALM apprenticeship at AFD is already the first link. ### Why the fit - **Apprenticeship secured**: ALM at AFD (two-year alternance from Sep 2026). - **Applied Data & AI**: Python pipelines, LLM orchestration, 93-node n8n workflows, Apex + LWC, scoring algorithms. Deployed across 6+ production projects for paying clients. Not coursework. - **Academic fit**: dual background in economics and computer science. Self-study in quant finance and ML, Bloomberg BMC in progress. - **Execution track record**: from scraper to dashboard to live bot. Not demos: running systems I rely on myself. - **Field-ready**: apprenticeship already secured at a public development bank. I arrive day one on an ALM desk, not in class. - **Skin in the game**: active investor, real money in ETFs and crypto, real estate deal underwriting on the side. Markets are my hobby, not my homework. --- ## Contact MSc questions, apprenticeship conversations, freelance briefs, or a quant problem over coffee. Pick the channel that fits. - **Email (best for detailed briefs)**: contact@mathieuhaye.fr - **LinkedIn (best for a quick intro)**: https://www.linkedin.com/in/mathieu-haye/ - **Phone (best for a fast call)**: +33 6 61 51 32 89 **Based in**: Paris, France **Languages**: French native, English B2 / C1 **Mobility**: Driver licence, Île-de-France **Status**: ALM apprentice, AFD - Sep 2026 --- *© 2026 Mathieu Haye. Hand-coded. HTML · CSS · JS · IONOS.*